Balance Sheet Normalization Roadmaps

The Federal Reserve's balance sheet peaked at $8.9 trillion in 2022, shrank to roughly $6.5 trillion by late 2025, and then QT stopped. That trajectory—$2.4 trillion drained in about three and a half years—reshaped Treasury supply dynamics, repo market plumbing, and term premia in ways most investors still underestimate. The practical point isn't predicting the next balance sheet move. It's understanding how reserve levels drive funding conditions so you can position ahead of liquidity regime shifts rather than reacting after spreads blow out.

Why Balance Sheet Size Matters to Your Portfolio (Not Just to Banks)

You might think the Fed's balance sheet is a concern for money market traders and bank treasurers—not for you. That's wrong, and the September 2019 repo crisis proved it.

Here's the causal chain:

Fed drains reserves (QT) → Bank reserves decline → Funding markets tighten → Repo rates spike → Treasury market liquidity deteriorates → Risk assets reprice

In September 2019, aggregate reserves had fallen to a multi-year low of less than $1.4 trillion after two years of balance sheet runoff. On September 16, two routine events collided: quarterly corporate tax payments drained cash from the banking system, and new Treasury securities settled (requiring dealer financing). The overnight repo rate spiked to 10%—from a normal range around 2%. The Fed had to intervene with emergency repo operations and eventually restart Treasury bill purchases at $60 billion per month.

What the data confirms: Reserve scarcity doesn't announce itself gradually. It builds quietly during balance sheet runoff, then erupts suddenly when a routine cash demand hits a system running on fumes. You don't get a polite warning—you get a funding crisis on a random Tuesday in September.

How QT Actually Worked (The Mechanics You Need to Know)

The Fed doesn't sell bonds to shrink its balance sheet. It simply stops reinvesting maturing securities—a process called "passive runoff." The FOMC sets monthly caps on how much can roll off, and anything above the cap gets reinvested.

The runoff cap timeline tells the whole story:

The point is: The Fed didn't run QT at a constant pace—it decelerated twice before stopping entirely. That pattern of progressive slowdown matters because it reveals the Fed's risk management approach. They'd rather stop too early than trigger another 2019-style blowup. If you were waiting for a dramatic "QT is over" announcement, you missed that the effective tightening had already slowed to a crawl months before the official end.

Why MBS runoff consistently undershot the cap: The $35 billion MBS cap was rarely binding. In a high-rate environment (with mortgage rates near 7%), almost no one refinances, so MBS prepayments slowed to a trickle. Actual MBS runoff ran closer to $15-20 billion per month—well below the cap. Over the entire QT period, MBS holdings declined by only about $600 billion versus roughly $1.6 trillion for Treasuries. The Fed wanted to shift toward a Treasury-only portfolio, but rate dynamics worked against them.

The practical takeaway: When you hear "the Fed is running off $95 billion per month," check the actual pace. Caps are ceilings, not floors. MBS prepayment dynamics mean actual shrinkage almost always runs below the announced maximum.

The Scarcity Signal You Should Have Been Watching

The Fed operates in what it calls an "ample reserves" regime. The goal is keeping enough reserves in the banking system that the fed funds rate stays within its target range without daily intervention. Too many reserves and you've got excess liquidity sloshing around. Too few and funding markets seize up (as in September 2019).

The critical question throughout QT was always: how do you know when reserves shift from "ample" to "scarce"?

Here are the signals the Fed watched—and that you should track:

Early warning indicators:

• EFFR drift: The effective fed funds rate moving persistently into the upper half of the target range (normally it sits 7-8 basis points above the floor)
• ON RRP decline: The overnight reverse repo facility dropping below $100 billion signals the easy liquidity cushion has been absorbed
• Month-end repo volatility: Spikes in secured overnight financing rate (SOFR) at quarter-end and month-end indicate dealer balance sheet constraints

What actually triggered the end of QT: By October 2025, money market rates began drifting higher. Utilization of the Fed's Standing Repo Facility—a backstop designed after the 2019 crisis—increased notably. These were the exact early-warning signals the Fed had identified. Rather than waiting for a full-blown crisis, the FOMC announced in late October 2025 that QT would end on December 1, 2025.

Why this matters: The Fed ended QT earlier than most market participants expected. The median respondent in the New York Fed's Survey of Market Expectations had projected QT ending in early 2026. If you were positioned for continued balance sheet runoff through Q1 2026, you got caught leaning the wrong way.

What the Final Balance Sheet Looks Like (And What Comes Next)

The Fed's balance sheet settled around $6.5 trillion when QT ended—down from $8.9 trillion at the peak but still roughly three times larger than pre-2020 levels (and massively larger than the pre-2008 balance sheet of about $900 billion).

The composition shift matters more than the total size:

Two things jump out. First, MBS still represent roughly a third of the portfolio—far more than the Fed wants long-term (they prefer an all-Treasury balance sheet to avoid influencing the housing market). Second, reserve balances didn't decline as dramatically as total assets because the ON RRP facility (which absorbed excess liquidity on the liability side) dropped from over $2.5 trillion to under $200 billion, absorbing much of the adjustment.

The lesson worth internalizing: Most of QT's "work" was done not by draining bank reserves but by draining the ON RRP facility—essentially unwinding the excess liquidity that money market funds had parked at the Fed. Reserves themselves fell only modestly, which is why QT didn't cause more disruption than it did. Once the ON RRP cushion was exhausted, further runoff would have directly drained reserves, which is precisely why the Fed stopped.

What Happens After QT Ends (The Post-Runoff Playbook)

Starting December 2025, the Fed began rolling over all maturing Treasuries and reinvesting MBS principal payments into Treasury bills. This isn't neutral—it's a deliberate shift. By converting MBS runoff into T-bill purchases, the Fed gradually moves its portfolio toward the all-Treasury composition it prefers (without actively selling MBS, which could disrupt mortgage markets).

Reserve Management Purchases (RMPs) are the next phase to watch. Over one-third of respondents in the NY Fed's market survey expected RMPs to be announced, with most anticipating net purchases of roughly $220 billion in the first twelve months. These aren't QE—they're designed to keep reserves growing in line with the economy's expanding demand for central bank liabilities (currency in circulation grows about $100-150 billion per year, and reserves need to grow proportionally).

The test: Watch whether the Fed characterizes any future purchases as "technical" (maintaining reserves) or "stimulative" (easing financial conditions). The framing matters enormously for market psychology—and for how aggressively you should position for lower yields.

How the ECB and BOJ Compare (And Why You Should Care)

Central bank balance sheet normalization isn't just a Fed story. The global liquidity picture depends on what all three major central banks do simultaneously.

ECB: The European Central Bank has been running its own normalization since 2023, ending reinvestments under the Asset Purchase Programme (APP) and subsequently the Pandemic Emergency Purchase Programme (PEPP) in 2024. The ECB's balance sheet has fallen to roughly 40% of eurozone GDP, and ECB Governing Council member Isabel Schnabel notably frames the process as "Quantitative Normalization" (QN)—not QT—emphasizing it as a return to normal rather than active tightening. The ECB expects a further €500 billion decline in non-borrowed reserves through 2025 as APP and PEPP holdings mature.

BOJ: The Bank of Japan is the outlier. After ending negative interest rates in March 2024 and raising rates to 0.25%, the BOJ is actively accelerating QT with a target of reducing holdings by ¥3 trillion per month by Q1 2026. The BOJ is the only major central bank simultaneously hiking rates and accelerating balance sheet reduction—a combination that has historically been volatile for JGB markets (and, by extension, for global bond flows given Japan's status as a major capital exporter).

The practical point: When all three major central banks drain reserves simultaneously, global liquidity tightens more than any single central bank's actions suggest. If you're pricing risk assets based only on Fed policy, you're missing half the picture. The BOJ's normalization, in particular, could redirect Japanese institutional capital flows (a perennial driver of Treasury demand) in ways that amplify term premium pressures on U.S. long bonds.

Term Premium and What Balance Sheet Policy Means for Bond Pricing

QT doesn't just drain reserves—it increases the supply of duration that private markets must absorb. When the Fed stops buying (or actively runs off) long-dated Treasuries and MBS, private investors demand compensation for holding that duration risk. This compensation shows up as term premium—the extra yield investors require for holding long-term bonds instead of rolling short-term instruments.

Through 2025, the long end of the curve remained constrained by persistent term-premium pressures. High term premia and weak long-end demand limited duration upside even as short rates were being cut.

The chain:

Fed stops buying long bonds → More duration for private markets to absorb → Term premium rises → Long yields stay elevated even as short rates fall → Yield curve steepens

This explains a puzzle many investors found confusing in 2024-2025: the Fed was cutting rates, but 10-year yields remained stubbornly high. Part of the answer is QT. By reducing its holdings of long-dated securities, the Fed was effectively pushing duration supply back into private hands—offsetting some of the easing from rate cuts with tightening from balance sheet policy.

The takeaway: Rate cuts and QT are partially contradictory policies. One eases financial conditions; the other tightens them. When you're modeling the net policy impulse, you have to account for both. An investor who saw "Fed is cutting rates" and piled into long duration without considering QT's term premium effect got an unpleasant surprise.

Detection Signals (How to Know the Next Normalization Cycle Is Coming)

You'll face another QT cycle eventually—possibly after the next recession triggers a new round of QE. Here's how to spot the normalization turning points:

QT is approaching if:

• The Fed signals "ample reserves" in FOMC minutes and begins discussing "appropriate balance sheet size"
• ON RRP usage is elevated (above $1 trillion), providing a cushion to absorb initial runoff
• The FOMC introduces runoff caps language in its post-meeting statement
• Fed officials begin speeches distinguishing between rate policy and balance sheet policy

QT is ending if:

• EFFR consistently trades in the upper half of the target range
• Standing Repo Facility usage increases beyond routine levels
• ON RRP falls below $200 billion (the liquidity cushion is gone)
• FOMC minutes show debate about "approaching" or "nearing" ample reserves
• Repo rate volatility spikes at quarter-end without obvious seasonal explanation

Monitoring Checklist (Tiered by Impact)

Essential (prevents being blindsided)

These four items give you 80% of the signal:

• Track the weekly H.4.1 release (every Thursday) for total securities, reserve balances, and ON RRP usage
• Monitor EFFR's position within the target range (watch for persistent drift toward the ceiling)
• Read the FOMC statement's balance sheet paragraph after every meeting (changes in language are leading indicators)
• Check SOFR volatility around month-end and quarter-end dates

High-impact (for active fixed-income positioning)

For investors managing duration or funding risk:

• Build a simple model tracking actual vs. cap runoff pace (actual pace reveals MBS prepayment dynamics)
• Monitor Standing Repo Facility and discount window usage via NY Fed data
• Track net Treasury issuance alongside Fed runoff to estimate private-sector absorption requirements
• Follow BOJ and ECB balance sheet trajectories for global liquidity context

Advanced (for macro-focused allocators)

If you're managing cross-asset positioning:

• Decompose 10-year yields into rate expectations and term premium (the NY Fed publishes an ACM model estimate)
• Track primary dealer Treasury inventory levels as a proxy for market absorption capacity
• Monitor Japanese life insurer and pension fund flows into U.S. Treasuries (affected by BOJ policy and hedging costs)

Your Next Step

Pull up the Fed's H.4.1 statistical release (published every Thursday at 4:30 PM ET on the Federal Reserve's website). Find three numbers: (1) total securities held outright, (2) reserve balances with Federal Reserve Banks, and (3) reverse repurchase agreements. Write them down. Do the same thing next Thursday. The week-over-week change in securities tells you how fast the balance sheet is actually moving. The reserve balance level tells you how much liquidity cushion remains. And the ON RRP level tells you whether the easy adjustment phase is over or still has room to run. After four weeks of this, you'll have better intuition for Fed plumbing than 90% of market commentators—and you'll recognize the early warning signals of reserve scarcity before they become front-page news.